|
Dear Carol:
Generally, you can compute a CI on the log(titer)
values and the exponentiate. I wrote some code to do this and also do
a boostrap CI, just for fun. Not the default dilution factor is 2, but you
may want to switch this to 10, say. The function will perform separate
analyses with different groups, specified with a formula; for one group, you
need to use a formula as Y ~ 1, or you can use the gmt.basic function
directly. Here's some example calls:
gmt.dat <- data.frame(Y2 =
2^c(0,2,5,3,1,4,4,3,2,2,1,2,3), Y10 =
10^c(1,0,0,2,1,1,3,1,2,2,1,1,3),
Grp=c("A","A","A","A","A","A","B","B","B","B","B","B","B")) gmt(Y2~1,data="">gmt(Y2~Grp,data="">gmt(Y10~Grp,data=""> Note that this is not designed to *compare* the groups,
which would be nice, huh?
Cheers,
Brad
Brad Biggerstaff, Ph.D. From: s-news-owner@lists.biostat.wustl.edu [mailto:s-news-owner@lists.biostat.wustl.edu] On Behalf Of carol white Sent: Friday, January 11, 2008 4:46 AM To: s-news@wubios.wustl.edu Subject: [S] CI of geometric mean Hi, It doesn't seem that the calculation of Confidence Interval of geometric mean is already implemented. Does any body have the function written in S? Thanks carol Looking for last minute shopping deals? Find them fast with Yahoo! Search.
|
